Bipolar transistors with an electrode control the surface potential on the emitter-base junction are fabricated, and the DC and 1/f noise characteristics of the transistors are evaluated. The 1/f noise in the transistors fabricated through the improved process were mainly generated at the bottom of the emitter regions. The contribution of the surface phenomena to the 1/f noise is less than 10% of the total 1/f noise. An equation is derived for the 1/f noise in bipolar transistors in terms of the recombination current in the emitter-base space charge regions. The equation is supported by the experimental results that the recombination-free transistor which are fabricated through the clean processes do not generate 1/f noise.

抄録全体を表示

The Internal Stress and Young's modulus changes of LPCVD silicon nitride thin film by ion implantation were measured. The load-deflection measurement technique of the square membrane with size of 2mm×2mm was used. The tensile internal stress σ was reduced from 1.2 GPa to less than 0.1GPa and the Young's modulus E was reduced from 320 GPa to 260 Gpa. The effects of σ and E of a pressure sensor diaphragm, with a side of 2a and a thickness of t, on pressure sensitivity were calculated. In the case of (t/a)²_??_σ/E, lowering of sensitivity due to σ can not be neglected. Therefore, it appears that the sensitivity of micro-pressure sensor with silicon nitride film as the construction material for diaphragm can be improved by several times by utilizing the ion implantation technique.

抄録全体を表示

This paper studies properties of limiting zeros, which are defined as zeros of sampled-data systems resulting from continuous-time scalar systems preceded by zero-order hold with sufficiently small sampling period. Firstly, the regions are clarified in which the limiting zeros corresponding to the zeros of the original continuous-time system should exist. Secondly, other limiting zeros are studied, which are always yielded when the relative degree of the original continuous-time system is greater than or equal to 2 and do not have any counterpart of the original continuous-time system. To be specific, stability of such a zero is studied for the case where the relative degree of the original continuous-time system is 2. Using the above two results, a condition is derived on which all the limiting zeros are stable.
Furthermore, relating to strong stabilizability of sampled-data systems, which is defined as possibility of being stabilized by a stable controller and is equivalent to the parity interlacing property (a condition on the location of real zeros and real poles), some criteria are given with which to examine whether limiting zeros are real or not.

抄録全体を表示

It is very important for the furnace design to estimate the load conductance between the electlodes.
A method of determining that of the 50 and 60Hz furnace is gven in this paper, as follows:
It is supposed for a pair of electrodes observed that one is applied to the effective voltage between them and the other zero, and the static electric field is analyzed by a finite element method and the heat flow generated in the charge of the furnace is calculated.
In the case of a two-electrode single phase furnace, the load conductance is obtained by dividing the heat flow calculated above by the square of the effective voltage. In the case of a three-electrode or six-electrode three-phase furnace, the generated heat flows for three pairs of electrodes are calculated as the above and three conductances are obtained by dividing them by the square of the effective voltage. The equations for three unkown load conductances between the electrodes are derived by using three conductances and solved.
A division method of the domain of the static electric field with the 2-order triangular isoparametric elements and a numerical method to estimate exactly the load conductance between electrodes of two-electrode-in-line single phase and three-electrode-in-line or six-electrode-in-line three-phase glass-melting furnaces are given. Computed results are shown.

抄録全体を表示

An absolute pressure sensor has reference cavity and elecrtical feedthrough from the cavity. The reference cavity has to be hermetically sealed, but the electrical feedthrough at the bonding surface makes small leakage.
In this paper, a new feedthrough structure is proposed and is applied to a piezoresistive pressure sensor.
A pyrex glass and a sillicon are used for this structure. The pyrex glass has fine through-holes engraved by electrolytic discharge drilling. The silicon has the diffused electrical feedthrough from the reference cavity. The reference cavity is hermetically sealed by glass-silicon anodic bonding. Bonding pads are made by Cr-Cu-Au mask evaporation to the glass holes. Lead wires are connected to the pads inside the glass hole with conductive epoxy or solder.
The advantages of this structure are as follows.
(1) Reference cavity is hermetically sealed without leakage by glass-silicon anodic bonding.
(2) Sensor is assembled in small size.
(3) Lead wires are attached at the opposite side of the diaphragm and strongly enough.
The piezoresistive absolute pressore sensor with this structure is small and high performance. This structure is also useful for capacitive absolute pressure sensors.

抄録全体を表示

The importance of SIS (Strategic Information System), which supports strategic business operation, has been increasing with the progress of information and communication technology. In SIS, the close connectivity between information and business is a critical success factor. In this paper the authors propose a methodology of planning SIS. First, introducing the business information system diagram, the role of information for business is clarified. Next, such concept is introduced that the information power as effectiveness for business is determined by the connectivity between information quality and information field representing the necessity of information. On the basis of this concept, a methodology of planning SIS, which consists of 3 factors, that is, steps, tools and frameworks, is proposed. The characteristic of the proposed methodology is to find a solution to the complex problem of planning SIS systematically by employing fixed steps, fixed tools and fixed frameworks.

抄録全体を表示

Several conditions are derived for a polytope of Matrices to belong to P-matrices or M-matrices. These matrix classes play a significant role in many fields of systems science including econometrics, numerical analysis, nonlinear network theory, control theory, and so on. Since a polytope of matrices is a typical way to represent matrices whose entries are uncertain, the results obtained will render themselves useful tools, when one tries to introduce the robustness concept into the above theories.
The first major result of the present paper is a sufficient condition, which is expressed as a simple constraint to each vertex matrix. This is in force both for P-matrices and M-matrices. However, much simpler condition is available especially for M-matrices. As the second result, necessary and sufficient conditions are given for a convex combination of two matrices to be in the two matrix classes. Thus, we can obtain exact conditions when the number of vertex matrix is only two. All these results can be carried over to interval matrices, because an interval matrix can be expressed as a polytope of matrices whose entries are either of the end points of given intervals. Some comments are made on the parallelism of the properties between a polytope of matrices and that of polynomials.

抄録全体を表示

In this paper, the strategy for estimation of rotor flux using the sliding observer, which is robust to the measurement noises, is proposed. Furthermore, the method of designing a robust sliding observer taking account of the ability of disturbance attenuation is also derived. Finally, the geometric structures of all the disturbances existing in the estimation problem are clarified, and the optimal sliding observer against those disturbances are derived

抄録全体を表示